Fuzzy rules may be expressed in terms such as ``If the room gets hotter, spin the fan blades faster'' where the temperature of the room and speed of the fan's blades are both imprecisely (fuzzily) defined quantities, and ``hotter'' and ``faster'' are both fuzzy terms. Fuzzy logic, with fuzzy rules, has the potential to add human-like subjective reasoning capabilities to machine intelligences, which are usually based on bivalent boolean logic. [Wang, 1993] provides an excellent introduction to fuzzy logic and how it can be applied to AI, and [Kosko, 1993] is an excellent popular science book describing all aspects of fuzzy logic.
For the most part, the fuzzy rules that are used in control systems are hand-crafted by the designers of the systems, and machine learning is rarely employed1. As such, it can be argued that this human input into the system's design constitutes a homunculus, and that such systems can never be independently intelligent2.
This project aimed to exploit this in order to investigate the effectiveness of machine learning algorithms when compared to human learning processes. The fuzzy rules served well as they provide an excellent means of compactly representing what the human has learned in a way that is both accurately human-like and subjective, and can be used in controlled experiments to be compared against standard machine learning algorithms.
The compactness of the rules is desirable because there exists evidence suggesting smaller rules perform better [Holte, 1993], with reasons essentially the same as those for overfitting in decision trees - the underlying structure of the process generating the data is captured rather than the superficial structure of the training data. Fuzzy logic appears to be very well suited to the creation of small rules, as fuzzy rules have a higher ``information density'', so to speak - each rule encapsulates a richness of information and meaning.